In [378]:
import matplotlib.pyplot as plt
%matplotlib inline
import os
import numpy as np
from skimage.exposure import equalize_adapthist
from skimage.transform import resize
from skimage.color import rgb2ycbcr
from scipy.misc import imsave
import cv2
import glob
# Few helper functions
def show_images(images,titles=None, save=None):
"""Display a list of images"""
n_ims = len(images)
if titles is None: titles = ['(%d)' % i for i in range(1,n_ims + 1)]
fig = plt.figure()
n = 1
for image,title in zip(images,titles):
a = fig.add_subplot(1,n_ims,n) # Make subplot
if len(image.shape) == 2 or image.shape[2] == 1: # Is image grayscale?
plt.imshow(np.resize(image, (image.shape[0], image.shape[1])), interpolation="bicubic", cmap="gray") # Only place in this blog you can't replace 'gray' with 'grey'
else:
plt.imshow(image, interpolation="bicubic")
if titles is not None:
a.set_title(title)
n += 1
fig.set_size_inches(np.array(fig.get_size_inches()) * n_ims)
plt.tight_layout()
plt.show()
if save is not None:
fig.savefig("examples/" + save + ".png")
def extract_frames(clip, times, imgdir, imgname):
for t in times:
imgpath = os.path.join(imgdir, '{}-{}.jpg'.format(imgname, t))
clip.save_frame(imgpath, t)
def draw_corners(img_fname, nx=9, ny=6):
img = cv2.imread(img_fname)
# Convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# Find the chessboard corners
ret, corners = cv2.findChessboardCorners(gray, (nx, ny), None)
# If found, draw corners
if ret == True:
# Draw and display the corners
return ret, corners, cv2.drawChessboardCorners(img, (nx, ny), corners, ret)
else:
return ret, corners, None
def output_image(img, img_name, prefix="examples"):
imsave(str(prefix) + "/" + str(img_name) + ".png", img)
In [2]:
show_images([draw_corners("camera_cal/calibration2.jpg")[2]])
In [3]:
def get_calibrated(images, nx=9, ny=6):
objpoints = [] # 3D points in the real world
imgpoints = [] # 2D points in the images
# Prepare object points for known corners
objp = np.zeros((nx*ny, 3), np.float32)
objp[:,:2] = np.mgrid[0:nx, 0:ny].T.reshape(-1, 2) # x, y coordinates
for img in images:
ret, corners, img_corners = draw_corners(img, nx=nx, ny=ny)
if ret == True:
objpoints.append(objp)
imgpoints.append(corners)
img = cv2.imread(images[0])
return cv2.calibrateCamera(objpoints, imgpoints, img.shape[1::-1], None, None)
images = glob.glob("camera_cal/calibration*.jpg")
# Leave 1st image as test image.
ret, mtx, dist, _, _ = get_calibrated(images[1:])
In [4]:
images_undist = [ cv2.undistort(cv2.imread(x), mtx, dist, None, mtx) for x in images ]
for i in range(0, 4):
show_images([cv2.imread(x) for x in images[i*5:i*5+6]], ["Distored"]*5)
show_images(images_undist[i*5:i*5+6], ["Un-Distored"]*5)
In [373]:
def plot_dist_undist(dist, undist,save_name):
fig, (ax1, ax2) = plt.subplots(2, 1)
ax1.imshow(dist)
ax1.set_title("distorted")
ax2.imshow(undist)
ax2.set_title("un-distorted")
plt.tight_layout()
plt.draw()
fig.savefig("examples/" + save_name + ".png")
plot_dist_undist(cv2.imread(images[9]), images_undist[9], "camera_dist")
In [5]:
from moviepy.editor import VideoFileClip
from IPython.display import HTML
def extract_images(video_fname, times, imgdir, imgname):
clip1 = VideoFileClip(video_fname)
extract_frames(clip1, times, imgdir, imgname)
In [110]:
# Extract images from videos
extract_images("project_video.mp4", [49.0, 50.0], "test_images", "project_video")
extract_images("challenge_video.mp4", [1.0, 2.0, 3.0, 3.5, 4.3, 4.4, 4.6, 4.8, 5.0], "test_images", "challenge_video")
extract_images("harder_challenge_video.mp4", [3.0, 4.0], "test_images", "harder_challenge_video")
In [111]:
# undistort test images
test_imgs = ["test_images/project_video-50.0.jpg", "test_images/challenge_video-2.0.jpg", "test_images/harder_challenge_video-3.0.jpg"]
undist_test_imgs = [cv2.cvtColor(cv2.undistort(cv2.imread(x, cv2.IMREAD_COLOR), mtx, dist, None, mtx), cv2.COLOR_BGR2RGB) for x in test_imgs]
show_images(undist_test_imgs)
In [112]:
# Prespective transform
bottom_left = (290, 673)
bottom_right = (1024, 673)
top_right = (757,496)
top_left = (531, 496)
src = np.float32([[bottom_left, bottom_right, top_right, top_left]])
dst = np.float32([[(460, 673), (820, 673), (820, 496), (460, 496)]])
M = cv2.getPerspectiveTransform(src, dst)
Minv = cv2.getPerspectiveTransform(dst, src)
img = undist_test_imgs[0]
print(img.shape)
def warp(img):
return cv2.warpPerspective(img, M, (img.shape[1], img.shape[0]), flags=cv2.INTER_LINEAR)
def unwarp(img):
return cv2.warpPerspective(img, Minv, (img.shape[1], img.shape[0]), flags=cv2.INTER_LINEAR)
warped = warp(img)
inv_warped = unwarp(warped)
f, (ax1, ax2, ax3) = plt.subplots(3, sharex=True)
ax1.imshow(img)
ax2.imshow(warped)
ax3.imshow(inv_warped)
plt.show()
f, ax = plt.subplots(1)
ax.imshow(warped)
plt.show()
In [380]:
# Thresholding
def get_hls(img_rgb):
return cv2.split(cv2.cvtColor(img_rgb, cv2.COLOR_RGB2HLS))
def get_rgb(img_rgb):
return cv2.split(img_rgb)
def show_spaces(img, save=None):
h, l, s = get_hls(img)
img_list = [img, get_rgb(img)[0], h, l, s]
name_list = ["orig", "red", "hue", "lighting", "saturation"]
show_images(img_list, name_list, save=save)
show_spaces(img, save="color_split")
show_spaces(undist_test_imgs[1])
show_spaces(undist_test_imgs[2])
In [381]:
def show_spaces(img, save=None):
h, l, s = get_hls(img)
img_list = [img, get_rgb(img)[0], s]
name_list = ["orig", "red", "saturation"]
show_images(img_list, name_list, save=save)
show_spaces(img, save="color_split")
In [383]:
clahe = cv2.createCLAHE()
def _rbg_CLAHE(img):
img_lab = cv2.cvtColor(img, cv2.COLOR_RGB2Lab)
l,a,b = cv2.split(img_lab)
l_clahe = clahe.apply(l)
return cv2.cvtColor(cv2.merge((l_clahe, a, b)), cv2.COLOR_Lab2RGB)
def abs_sobel_thresh(img, orient='x', sobel_kernel=5, thresh=(0, 255), is_gray=True):
if is_gray:
gray = img
else:
gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
if orient == 'x':
sobel = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=sobel_kernel)
else:
sobel = cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=sobel_kernel)
abs_sobel = np.absolute(sobel)
scaled_sobel = np.uint8(255*abs_sobel/np.max(abs_sobel))
sbinary = np.zeros_like(scaled_sobel)
sbinary[(scaled_sobel >= thresh[0]) & (scaled_sobel <= thresh[1])] = 1
return sbinary
def dir_threshold(img, sobel_kernel=5, thresh=(0, np.pi/2), is_gray=True):
# Grayscale
if is_gray:
gray = img
else:
gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
# Calculate the x and y gradients
sobelx = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=sobel_kernel)
sobely = cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=sobel_kernel)
# Take the absolute value of the gradient direction,
# apply a threshold, and create a binary image result
absgraddir = np.arctan2(np.absolute(sobely), np.absolute(sobelx))
binary_output = np.zeros_like(absgraddir)
binary_output[(absgraddir >= thresh[0]) & (absgraddir <= thresh[1])] = 1
return binary_output
def mag_thresh(img, sobel_kernel=5, mag_thresh=(0, 255), is_gray=True):
# Apply the following steps to img
# 1) Convert to grayscale
if is_gray:
gray = img
else:
gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
# Take both Sobel x and y gradients
sobelx = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=sobel_kernel)
sobely = cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=sobel_kernel)
# Calculate the gradient magnitude
gradmag = np.sqrt(sobelx**2 + sobely**2)
# Rescale to 8 bit
scale_factor = np.max(gradmag)/255
gradmag = (gradmag/scale_factor).astype(np.uint8)
# Create a binary image of ones where threshold is met, zeros otherwise
binary_output = np.zeros_like(gradmag)
binary_output[(gradmag >= mag_thresh[0]) & (gradmag <= mag_thresh[1])] = 1
return binary_output
def combine_thresh(image, ksize=5, is_gray=True):
# Apply each of the thresholding functions
gradx = abs_sobel_thresh(image, orient='x', sobel_kernel=ksize, thresh=(20, 150), is_gray=is_gray)
grady = abs_sobel_thresh(image, orient='y', sobel_kernel=ksize, thresh=(20, 150), is_gray=is_gray)
mag_binary = mag_thresh(image, sobel_kernel=ksize, mag_thresh=(20, 150), is_gray=is_gray)
dir_binary = dir_threshold(image, sobel_kernel=ksize*3, thresh=(0.7, 1.3), is_gray=is_gray)
combined = np.zeros_like(dir_binary)
combined[((gradx == 1) & (grady == 1)) | ((mag_binary == 1) & (dir_binary == 1))] = 1
return combined, gradx, grady, mag_binary, dir_binary
def show_thresh(img, apply_clahe=False):
if apply_clahe:
clahe = _rbg_CLAHE(img)
else:
clahe = img
r, g, b = get_rgb(clahe)
h, l, s = get_hls(clahe)
show_images([img, combine_thresh(r)[0], combine_thresh(h)[0], combine_thresh(l)[0], combine_thresh(s)[0]], ["orig", "red", "hue", "lighting", "saturation"])
show_images([img, mag_thresh(get_hls(img)[2], sobel_kernel=5, mag_thresh=(20, 150))], ["orig", "mag_thresh"])
show_thresh(img)
show_thresh(undist_test_imgs[1])
show_thresh(undist_test_imgs[2])
In [387]:
r_img = get_rgb(img)[0]
l = [r_img]
l.extend(combine_thresh(r_img)[1:])
show_images(l, ["red", "grad-x", "grad-y", "grad-mag", "grad-direction"], save="red_grads")
s_img = get_hls(img)[2]
l = [s_img]
l.extend(combine_thresh(s_img)[1:])
show_images(l, ["saturation", "grad-x", "grad-y", "grad-mag", "grad-direction"], save="sat_grads")
In [395]:
def combine_rs(img, apply_clahe=False):
"""
Expects a color image.
"""
if apply_clahe:
clahe = _rbg_CLAHE(img)
else:
clahe = img
r, g, b = get_rgb(clahe)
h, l, s = get_hls(clahe)
combined = np.zeros_like(h)
r_t = combine_thresh(r)[0]
s_t = combine_thresh(s)[0]
combined[((r_t == 1) | (s_t == 1))] = 1
return combined
def region_of_interest(img, vertices, mask_color = 255):
#defining a blank mask to start with
if mask_color == 255:
mask = np.zeros_like(img)
else:
mask = np.copy(img)
ignore_mask_color = mask_color
# filling pixels inside the polygon defined by "vertices" with the fill color
p = cv2.fillPoly(mask, vertices, ignore_mask_color)
# returning the image only where mask pixels are nonzero
masked_image = cv2.bitwise_and(img, p)
return masked_image
def lane_region_mask(image):
# copy the given arg so we do not mess with it.
# Grab the x and y size and make a copy of the image
lratio = 0.10
rratio = 0.04
ysize = image.shape[0]
xsize = image.shape[1]
# Mask the region of interest
region_array = np.array([[(int(xsize*lratio), image.shape[0]-0),
(int(xsize*(1.0-rratio)), image.shape[0]-0),
((xsize/2)*(1 + rratio*4), int(ysize*0.50)),
((xsize/2)*(1 - lratio*2), int(ysize*0.50))]],
dtype=np.int32)
region_array2 = np.array([[(int(xsize*lratio*3), image.shape[0]-0),
(int(xsize*(1.0-rratio*3.5)), image.shape[0]-0),
((xsize/2)*(1 + rratio*4), int(ysize*0.80)),
((xsize/2)*(1 - lratio*1), int(ysize*0.80))]],
dtype=np.int32)
return region_of_interest(region_of_interest(image, region_array), region_array2, mask_color = 0)
def warp_region(img, ignore_region=False):
# Mask the region of interest and feed through hough.
if ignore_region:
return warp(img)
else:
return warp(lane_region_mask(img))
def show_combine_rs(img, apply_clahe=False):
c = combine_rs(img, apply_clahe=apply_clahe)
show_images([img, c, lane_region_mask(c), warp(c), warp_region(c)], ["orig", "combined_l_s", "region", "warped", "warped_region"])
show_combine_rs(img)
show_combine_rs(undist_test_imgs[1])
show_combine_rs(undist_test_imgs[2])
#show_combine_ls(undist_test_imgs[3])
In [396]:
crs = combine_rs(img)
show_images([img, crs], ["orig", "combined red-saturation gradients"], save="comb_rs_grads")
crs_region = lane_region_mask(crs)
show_images([crs, crs_region], ["combined red-saturation gradients", "apply region"], save="comb_rs_grads_region")
In [415]:
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
def get_curvature(leftx, lefty, rightx, righty):
ym_per_pix = 30/720 # meters per pixel in y dimension
xm_per_pix = 3.7/700 # meters per pixel in x dimension
# Fit new polynomials to x,y in world space
left_fit_cr = np.polyfit(lefty*ym_per_pix, leftx*xm_per_pix, 2)
right_fit_cr = np.polyfit(righty*ym_per_pix, rightx*xm_per_pix, 2)
# Calculate the new radii of curvature
left_curverad = ((1 + (2*left_fit_cr[0]*np.max(lefty)*ym_per_pix + left_fit_cr[1])**2)**1.5) / np.absolute(2*left_fit_cr[0])
right_curverad = ((1 + (2*right_fit_cr[0]*np.max(righty)*ym_per_pix + right_fit_cr[1])**2)**1.5) / np.absolute(2*right_fit_cr[0])
# Now our radius of curvature is in meters
return left_curverad, right_curverad
def get_center_offset(left_fit, right_fit, ymax=720.0, xmax=1280.0):
xm_per_pix = 3.7/700 # meters per pixel in x dimension
left = left_fit[0]*ymax**2 + left_fit[1]*ymax + left_fit[2]
right = right_fit[0]*ymax**2 + right_fit[1]*ymax + right_fit[2]
return ((left + (right - left) / 2) - xmax/2)*xm_per_pix
# Sliding window
def sliding_window(binary_warped, window_img_out=False):
# Assuming you have created a warped binary image called "binary_warped"
# Take a histogram of the bottom half of the image
histogram = np.sum(binary_warped[int(binary_warped.shape[0]//2):,:], axis=0)
# Find the peak of the left and right halves of the histogram
# These will be the starting point for the left and right lines
midpoint = np.int(histogram.shape[0]/2)
leftx_base = np.argmax(histogram[:midpoint])
rightx_base = np.argmax(histogram[midpoint:]) + midpoint
# Choose the number of sliding windows
nwindows = 9
# Set height of windows
window_height = np.int(binary_warped.shape[0]/nwindows)
# Identify the x and y positions of all nonzero pixels in the image
nonzero = binary_warped.nonzero()
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
# Current positions to be updated for each window
leftx_current = leftx_base
rightx_current = rightx_base
# Set the width of the windows +/- margin
margin = 50
# Set minimum number of pixels found to recenter window
minpix = 35
# Create empty lists to receive left and right lane pixel indices
left_lane_inds = []
right_lane_inds = []
# Step through the windows one by one
if window_img_out:
out_img = np.dstack((binary_warped, binary_warped, binary_warped))*255
else:
out_img = None
for window in range(nwindows):
# Identify window boundaries in x and y (and right and left)
win_y_low = binary_warped.shape[0] - (window+1)*window_height
win_y_high = binary_warped.shape[0] - window*window_height
win_xleft_low = leftx_current - margin
win_xleft_high = leftx_current + margin
win_xright_low = rightx_current - margin
win_xright_high = rightx_current + margin
# Draw the windows on the visualization image
if window_img_out:
cv2.rectangle(out_img,(win_xleft_low,win_y_low),(win_xleft_high,win_y_high), (0,255,0), 2)
cv2.rectangle(out_img,(win_xright_low,win_y_low),(win_xright_high,win_y_high), (0,255,0), 2)
# Identify the nonzero pixels in x and y within the window
good_left_inds = ((nonzeroy >= win_y_low) & (nonzeroy < win_y_high) &
(nonzerox >= win_xleft_low) & (nonzerox < win_xleft_high)).nonzero()[0]
good_right_inds = ((nonzeroy >= win_y_low) & (nonzeroy < win_y_high) &
(nonzerox >= win_xright_low) & (nonzerox < win_xright_high)).nonzero()[0]
# Append these indices to the lists
left_lane_inds.append(good_left_inds)
right_lane_inds.append(good_right_inds)
# If you found > minpix pixels, recenter next window on their mean position
if len(good_left_inds) > minpix:
leftx_current = np.int(np.mean(nonzerox[good_left_inds]))
if len(good_right_inds) > minpix:
rightx_current = np.int(np.mean(nonzerox[good_right_inds]))
# Concatenate the arrays of indices
left_lane_inds = np.concatenate(left_lane_inds)
right_lane_inds = np.concatenate(right_lane_inds)
# Extract left and right line pixel positions
leftx = nonzerox[left_lane_inds]
lefty = nonzeroy[left_lane_inds]
rightx = nonzerox[right_lane_inds]
righty = nonzeroy[right_lane_inds]
# Fit a second order polynomial to each
left_fit, left_cov = np.polyfit(lefty, leftx, 2, cov=True)
right_fit, right_cov = np.polyfit(righty, rightx, 2, cov=True)
left_curverad, right_curverad = get_curvature(leftx, lefty, rightx, righty)
center_offset = get_center_offset(left_fit, right_fit)
return out_img, left_fit, right_fit, np.sqrt(np.diag(left_cov)), np.sqrt(np.diag(right_cov)), left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy
def visualize_silding_window(binary_warped, out_img, left_fit, right_fit, left_lane_inds, right_lane_inds, nonzerox, nonzeroy):
# Generate x and y values for plotting
# Create an output image to draw on and visualize the result
ploty = np.linspace(0, binary_warped.shape[0]-1, binary_warped.shape[0] )
left_fitx = left_fit[0]*ploty**2 + left_fit[1]*ploty + left_fit[2]
right_fitx = right_fit[0]*ploty**2 + right_fit[1]*ploty + right_fit[2]
out_img[nonzeroy[left_lane_inds], nonzerox[left_lane_inds]] = [255, 0, 0]
out_img[nonzeroy[right_lane_inds], nonzerox[right_lane_inds]] = [0, 0, 255]
f, ax = plt.subplots()
ax.imshow(out_img)
ax.plot(left_fitx, ploty, color='yellow')
ax.plot(right_fitx, ploty, color='yellow')
ax.set_xlim(0, 1280)
ax.set_ylim(720, 0)
plt.show()
def find_lanes_with_hint(binary_warped, left_fit, right_fit):
# Assume you now have a new warped binary image
# from the next frame of video (also called "binary_warped")
# It's now much easier to find line pixels!
nonzero = binary_warped.nonzero()
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
margin = 50
left_lane_inds = ((nonzerox > (left_fit[0]*(nonzeroy**2) + left_fit[1]*nonzeroy +
left_fit[2] - margin)) & (nonzerox < (left_fit[0]*(nonzeroy**2) +
left_fit[1]*nonzeroy + left_fit[2] + margin)))
right_lane_inds = ((nonzerox > (right_fit[0]*(nonzeroy**2) + right_fit[1]*nonzeroy +
right_fit[2] - margin)) & (nonzerox < (right_fit[0]*(nonzeroy**2) +
right_fit[1]*nonzeroy + right_fit[2] + margin)))
# Again, extract left and right line pixel positions
leftx = nonzerox[left_lane_inds]
lefty = nonzeroy[left_lane_inds]
rightx = nonzerox[right_lane_inds]
righty = nonzeroy[right_lane_inds]
# Fit a second order polynomial to each
if lefty.size == 0 or righty.size == 0:
return [None]*10
left_fit, left_cov = np.polyfit(lefty, leftx, 2, cov=True)
right_fit, right_cov = np.polyfit(righty, rightx, 2, cov=True)
left_curverad, right_curverad = get_curvature(leftx, lefty, rightx, righty)
center_offset = get_center_offset(left_fit, right_fit)
return left_fit, right_fit, np.sqrt(np.diag(left_cov)), np.sqrt(np.diag(right_cov)), left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy
def visualize_lanes(binary_warped, left_fit, right_fit, left_lane_inds, right_lane_inds, nonzerox, nonzeroy):
margin = 100
# Create an image to draw on and an image to show the selection window
out_img = np.dstack((binary_warped, binary_warped, binary_warped))*255
# Generate x and y values for plotting
ploty = np.linspace(0, binary_warped.shape[0]-1, binary_warped.shape[0] )
left_fitx = left_fit[0]*ploty**2 + left_fit[1]*ploty + left_fit[2]
right_fitx = right_fit[0]*ploty**2 + right_fit[1]*ploty + right_fit[2]
window_img = np.zeros_like(out_img)
# Color in left and right line pixels
out_img[nonzeroy[left_lane_inds], nonzerox[left_lane_inds]] = [255, 0, 0]
out_img[nonzeroy[right_lane_inds], nonzerox[right_lane_inds]] = [0, 0, 255]
# Generate a polygon to illustrate the search window area
# And recast the x and y points into usable format for cv2.fillPoly()
left_line_window1 = np.array([np.transpose(np.vstack([left_fitx-margin, ploty]))])
left_line_window2 = np.array([np.flipud(np.transpose(np.vstack([left_fitx+margin,
ploty])))])
left_line_pts = np.hstack((left_line_window1, left_line_window2))
right_line_window1 = np.array([np.transpose(np.vstack([right_fitx-margin, ploty]))])
right_line_window2 = np.array([np.flipud(np.transpose(np.vstack([right_fitx+margin,
ploty])))])
right_line_pts = np.hstack((right_line_window1, right_line_window2))
# Draw the lane onto the warped blank image
cv2.fillPoly(window_img, np.int_([left_line_pts]), (0,255, 0))
cv2.fillPoly(window_img, np.int_([right_line_pts]), (0,255, 0))
result = cv2.addWeighted(out_img, 1, window_img, 0.3, 0)
return result, left_fitx, right_fitx, ploty
def draw_visualize_lanes(binary_warped, left_fit, right_fit, left_lane_inds, right_lane_inds, nonzerox, nonzeroy):
result, left_fitx, right_fitx, ploty = visualize_lanes(binary_warped, left_fit, right_fit, left_lane_inds, right_lane_inds, nonzerox, nonzeroy)
f, ax = plt.subplots()
canvas = FigureCanvas(f)
ax.imshow(result)
ax.plot(left_fitx, ploty, color='yellow')
ax.plot(right_fitx, ploty, color='yellow')
ax.set_xlim(0, 1280)
ax.set_ylim(720, 0)
plt.show()
def draw_on_orig(undist, warped, left_fit, right_fit, text=None):
# Create an image to draw the lines on
warp_zero = np.zeros_like(warped).astype(np.uint8)
color_warp = np.dstack((warp_zero, warp_zero, warp_zero))
# Generate x and y values for plotting
ploty = np.linspace(0, warped.shape[0]-1, warped.shape[0])
left_fitx = left_fit[0]*ploty**2 + left_fit[1]*ploty + left_fit[2]
right_fitx = right_fit[0]*ploty**2 + right_fit[1]*ploty + right_fit[2]
# Recast the x and y points into usable format for cv2.fillPoly()
pts_left = np.array([np.transpose(np.vstack([left_fitx, ploty]))])
pts_right = np.array([np.flipud(np.transpose(np.vstack([right_fitx, ploty])))])
pts = np.hstack((pts_left, pts_right))
# Draw the lane onto the warped blank image
cv2.fillPoly(color_warp, np.int_([pts]), (0,255, 0))
# Warp the blank back to original image space using inverse perspective matrix (Minv)
newwarp = cv2.warpPerspective(color_warp, Minv, (undist.shape[1], undist.shape[0]))
# Combine the result with the original image
result = cv2.addWeighted(undist, 1, newwarp, 0.3, 0)
if text is not None:
font = cv2.FONT_HERSHEY_SIMPLEX
cv2.putText(result,text,(100,100), font, 1,(255,255,255),2,cv2.LINE_AA)
return result
def show_draw_on_orig(undist, warped, left_fit, right_fit, text=None):
result = draw_on_orig(undist, warped, left_fit, right_fit, text=text)
f, ax = plt.subplots()
ax.imshow(result)
ax.set_title(text)
plt.show()
def show_sliding_window(img, apply_clahe=False):
c = combine_ls(img, apply_clahe=apply_clahe)
warped = warp_region(c)
show_images([img, c, warped], ["orig", "combined_l_s", "warped_region", "sliding_window"])
f, ax = plt.subplots(1)
histogram = np.sum(warped[int(img.shape[0]/2):,:], axis=0)
ax.plot(histogram)
plt.show()
print(warped.shape)
out_img, left_fit, right_fit, left_conf, right_conf, left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy = sliding_window(warped, window_img_out=True)
print("left conf: {} right conf: {}".format(left_conf, right_conf))
visualize_silding_window(warped, out_img, left_fit, right_fit, left_lane_inds, right_lane_inds, nonzerox, nonzeroy)
# find lane with hints, use the same img to make sure it is working
left_fit, right_fit, left_conf, right_conf, left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy = find_lanes_with_hint(warped, left_fit, right_fit)
draw_visualize_lanes(warped, left_fit, right_fit, left_lane_inds, right_lane_inds, nonzerox, nonzeroy)
text = "left {}m, right {}m, offset {:06.2f}m".format(int(left_curverad), int(right_curverad), center_offset)
show_draw_on_orig(img, warped, left_fit, right_fit, text=text)
show_sliding_window(img)
In [419]:
warped = warp(crs_region)
out_img, left_fit, right_fit, left_conf, right_conf, left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy = sliding_window(warped, window_img_out=True)
left_fit, right_fit, left_conf, right_conf, left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy = find_lanes_with_hint(warped, left_fit, right_fit)
warped_lanes = visualize_lanes(warped, left_fit, right_fit, left_lane_inds, right_lane_inds, nonzerox, nonzeroy)[0]
text = "left {}m, right {}m, offset {:06.2f}m".format(int(left_curverad), int(right_curverad), center_offset)
orig_lanes = draw_on_orig(img, warped, left_fit, right_fit, text=text)
show_images([warped, warped_lanes, orig_lanes], ["birds eye", "lane detection", "unwrap lanes to origin"], save="bird_to_undist")
In [420]:
from collections import deque
class Line():
def __init__(self, max_len=15, max_miss=13, name=None):
self.max_len = int(max_len)
self.max_miss = int(max_miss)
self.name = name
self.miss_count = 0
self.line_q = deque()
self.line_sum = None
def is_valid(self):
return len(self.line_q) > 0
def get_line(self):
if len(self.line_q) == 0:
return None
else:
#print ("get_line[{}]: {}".format(self.name, self.line_sum / len(self.line_q)))
return self.line_sum / len(self.line_q)
def add_fit(self, line_fit):
self.line_q.append(line_fit)
if self.line_sum is None:
self.line_sum = np.copy(line_fit)
else:
self.line_sum += line_fit
self.miss_count = 0
if len(self.line_q) > self.max_len:
self._pop_line()
def add_unfit(self, line_fit):
self.miss_count += 1
if self.miss_count >= self.max_miss:
self._reset()
def _reset(self):
self.line_q.clear()
self.line_sum = None
self.miss_count = 0
def _pop_line(self):
e = self.line_q.popleft()
#print ("pop_line[{}]: sum {}, e {}".format(self.name, self.line_sum, e))
self.line_sum -= e
class Lanes():
def __init__(self, max_len=15, err_thresh=0.95):
self.max_len = int(max_len)
self.err_thresh = float(err_thresh)
self.left_fit = Line(max_len=max_len, name="left")
self.right_fit = Line(max_len=max_len, name="right")
self.left_curv = Line(max_len=max_len, name="lcurv")
self.right_curv = Line(max_len=max_len, name="rcurv")
def is_valid(self):
return self.left_fit.is_valid()
def reset(self):
self.left_fit._reset()
self.right_fit._reset()
self.left_curv._reset()
self.right_curv._reset()
def get_left_fit(self):
return self.left_fit.get_line()
def get_right_fit(self):
return self.right_fit.get_line()
def get_left_curv(self):
return self.left_curv.get_line()
def get_right_curv(self):
return self.right_curv.get_line()
def set_fit(self, left_fit, right_fit, left_curv, right_curv, left_err, right_err):
# print("left err: {}, right err: {}".format(left_err, right_err))
if np.all(left_err < self.err_thresh) and np.all(right_err < self.err_thresh):
self.left_fit.add_fit(left_fit)
self.right_fit.add_fit(right_fit)
self.left_curv.add_fit(left_curv)
self.right_curv.add_fit(right_curv)
else:
# print("left err: {}, right err: {}".format(left_err < self.err_thresh, right_err < self.err_thresh))
self.left_fit.add_unfit(left_fit)
self.right_fit.add_unfit(right_fit)
self.left_curv.add_unfit(left_curv)
self.right_curv.add_unfit(right_curv)
In [421]:
# Import everything needed to edit/save/watch video clips
from moviepy.editor import VideoFileClip
from IPython.display import HTML
def pipeline(img, mtx, dist, M, Minv, lanes_obj=None, apply_clahe=False):
"""
Undistort the image, threshold and apply prespective transform, find lanes,
then apply lane pixels to the image, re-distort the image and return it.
"""
# Undistort the image
undist = cv2.undistort(img, mtx, dist, None, mtx)
# Threshold application
c = combine_rs(img, apply_clahe=apply_clahe)
# Prespective transform and apply region of interest.
warped = warp_region(c)
# Apply sliding window to find lane lines
# TODO: Optimize this.
if lanes_obj is None or lanes_obj.is_valid() is False:
_, left_fit, right_fit, left_err, right_err, left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy = sliding_window(warped)
if lanes_obj is not None:
lanes_obj.set_fit(left_fit, right_fit, left_curverad, right_curverad, 0.0, 0.0)
else:
if lanes_obj.is_valid() is False:
print("left fit is none")
raise Exception("left fit is none")
left_fit, right_fit, left_err, right_err, left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy = find_lanes_with_hint(warped,
lanes_obj.get_left_fit(), lanes_obj.get_right_fit())
if left_fit is not None:
lanes_obj.set_fit(left_fit, right_fit, left_curverad, right_curverad, left_err, right_err)
if lanes_obj.is_valid() is False:
_, left_fit, right_fit, left_err, right_err, left_curverad, right_curverad, center_offset,left_lane_inds, right_lane_inds, nonzerox, nonzeroy = sliding_window(warped)
lanes_obj.set_fit(left_fit, right_fit, left_curverad, right_curverad, 0.0, 0.0)
else:
lanes_obj.reset()
_, left_fit, right_fit, left_err, right_err, left_curverad, right_curverad, center_offset, left_lane_inds, right_lane_inds, nonzerox, nonzeroy = sliding_window(warped)
lanes_obj.set_fit(left_fit, right_fit, left_curverad, right_curverad, 0.0, 0.0)
# light up the lane pixels
if lanes_obj.is_valid() is False:
return undist
else:
center_offset = get_center_offset(lanes_obj.get_left_fit(), lanes_obj.get_right_fit())
text = "left {}m, right {}m curvature, center offset {:06.2f}m".format(int(lanes_obj.get_left_curv()), int(lanes_obj.get_right_curv()), center_offset)
return draw_on_orig(undist, warped, lanes_obj.get_left_fit(), lanes_obj.get_right_fit(), text=text)
In [422]:
def process_video(input_video, output_video, clip=None):
white_output = 'output_videos/%s' % output_video
## To speed up the testing process you may want to try your pipeline on a shorter subclip of the video
## To do so add .subclip(start_second,end_second) to the end of the line below
## Where start_second and end_second are integer values representing the start and end of the subclip
## You may also uncomment the following line for a subclip of the first 5 seconds
if clip is not None:
clip1 = VideoFileClip('%s' % input_video).subclip(clip[0],clip[1])
else:
clip1 = VideoFileClip('%s' % input_video)
lanes = Lanes()
def process_image(img):
return pipeline(img, mtx, dist, M, Minv, lanes_obj=lanes)
white_clip = clip1.fl_image(process_image) #NOTE: this function expects color images!!
%time white_clip.write_videofile(white_output, audio=False)
return white_output
In [423]:
extract_images("project_video.mp4", [21.0, 22.0, 23.0], "test_images", "project_video")
lanes_try = Lanes()
def process_image(img):
return pipeline(img, mtx, dist, M, Minv, lanes_obj=lanes_try)
l = ["test_images/project_video-21.0.jpg", "test_images/project_video-22.0.jpg", "test_images/project_video-23.0.jpg"]
show_images([process_image(cv2.cvtColor(cv2.imread(l[0], cv2.IMREAD_COLOR), cv2.COLOR_BGR2RGB))], ["first"])
show_images([process_image(cv2.cvtColor(cv2.imread(l[1], cv2.IMREAD_COLOR), cv2.COLOR_BGR2RGB))], ["second"])
show_images([process_image(cv2.cvtColor(cv2.imread(l[2], cv2.IMREAD_COLOR), cv2.COLOR_BGR2RGB))], ["third"])
In [424]:
white_output = process_video("project_video.mp4", "project_video.mp4")
In [425]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(white_output))
Out[425]:
In [ ]: